F50 Predicting clinical scores in Huntington's disease: a lightweight speech test. (12th September 2022)

• Record Type: Journal Article
• Authors: Riad, Rachid; Lunven, Marine; Titeux, Hadrien; Hamet Bagnou, Jennifer; Lemoine, Laurie; Montillot, Justine; Sliwinski, Agnes; Youssov, Katia; Dupoux, Emmanuel; Bachoud-Lévi, Anne-Catherine
• Is Part Of: Journal of neurology, neurosurgery and psychiatry. Volume 93(2022)Supplement 1
• Page Start: A54
• Page End: A54
• Publication Date: 2022-09-12
• Subjects: Huntington's disease -- speech -- machine learning -- markers; Neurology -- Periodicals; Nervous system -- Surgery -- Periodicals; Psychiatry -- Periodicals; 616.8
• DOI: 10.1136/jnnp-2022-ehdn.141
• Languages: English
• ISSNs: 0022-3050

Abstract

Abstract : Using brief samples of speech recordings, we aimed at predicting, through machine learning, the clinical performance of Huntington's Disease (HD), an inherited neurodegenerative disease. We collected and analyzed 126 samples of audio recordings of both forward and backward counting from 103 Huntington's disease gene carriers [87 manifest and 16 premanifest; mean age 50.6 (SD 11.2), range (27–88) years] from three multicenter prospective studies in France and Belgium (MIG-HD BIO-HD and Repair-HD). We pre-registered our methods before any analyses to avoid inflated results. We automatically extracted 60 speech features from blindly annotated samples. We used machine learning models to combine speech features in order to make predictions at individual levels of the clinical markers. We trained models on 86% of the samples, the remaining 14% constituted the independent test set. We combined speech features with demographic variables (age, sex, CAG repeats, and burden score) to predict cognitive, motor, and functional scores of the Unified Huntington's disease rating scale. We provided correlation between speech variables and striatal volumes. Speech features combined with demographics allowed the prediction of the individual cognitive, motor, and functional scores with a relative error from 12.7 to 20.0% which is better than predictions using demographics and genetic information. Both mean and standard deviation of pause durations during backward recitation correlated with striatal atrophy (Spearman 0.6 and 0.5–0.6, respectively). Brief and examiner-free speech recording may become an efficient method for remote evaluation of the individual condition in HD.